dumplisp

[Andrew L. M. Shalit <alms>]

   Date: Mon, 4 Feb 91 10:22:45 CST
   From: lynch@aristotle.ils.nwu.edu (Richard Lynch)

   >I imagine it would be easy to explicitly code this behavior into
   >the functions which you add to *RESTORE-LISP-FUNCTIONS*.  You would
   >also have to backpatch the function(s) which are already on
   >*RESTORE-LISP-FUNCTIONS* (by wrapping them in another function
   >which calls them conditionally).
   >
   >   -andrew
   >
   >
   But what could the test be?  Any global var that is set would be STORED set
   in the image, hence re-starting the image would not work because the flag
   is stuck on...
   The only solution I can think of is to have a file with a single flag in it
   that tells whether or not we are in the process of quitting.  Seems
   excessive.

The only thing that comes to mind would be a check based on time.
You write an exit function which stores the time in a variable.
The restore functions only execute if a certain amount of time
has elapsed since the exit.

This won't be very portable, but you could probably get it to
work pretty easily in most cases.

   -andrew